Prakash Kondibhau Naikade · Thomas B. Moeslund · Andreas Møgelmose
AI:Xpertise Lab · Visual Analysis and Perception Laboratory, Aalborg University · Pioneer Centre for Artificial Intelligence
Project page · arXiv · Paper
BIMScript reconstructs an indoor scene as a short program of parametric commands in which every element carries not only geometry but the material and condition that define a building model — and each command maps one-to-one onto a native Revit/IFC object.
- What — per-element
material=andcondition=attributes, supervised by a VLM-distilled material passport over 1.9M elements across ~100k scenes. - Fast — decoding is bound by overhead, not compute: an output-exact CUDA-graph decoder cuts 6.40 → 1.91 ms/step (3.4×), composed with grammar-parallel draft-and-verify.
- Exactly where — a bounded sub-bin offset head lifts coordinates off the 5 cm token grid.
Code release is in preparation. Planned repositories in the BIMScriptWorld organization:
| Repository | Contents | Status |
|---|---|---|
BIMScript |
training, evaluation and decoding | coming soon |
bimscript-passport |
material-passport extraction pipeline and corpus | coming soon |
bimscript-revit |
pyRevit add-in and IFC4 export | coming soon |
Source lives in docs/ and is served by GitHub Pages at
https://BIMScriptWorld.github.io/BIMScript/. To preview locally:
cd docs && python -m http.server 8000 # then open http://localhost:8000@article{naikade2026bimscript,
title = {BIMScript: Material-Aware Structured Scene Programs for BIM Ingestion},
author = {Naikade, Prakash Kondibhau and Moeslund, Thomas B. and M{\o}gelmose, Andreas},
journal = {arXiv preprint arXiv:ARXIV_ID},
year = {2026}
}